電気学会論文誌C(電子・情報・システム部門誌)
Online ISSN : 1348-8155
Print ISSN : 0385-4221
ISSN-L : 0385-4221
117 巻, 2 号
選択された号の論文の16件中1~16を表示しています
  • 前田 裕, 四元 雄一郎, 金田 彌吉
    1997 年 117 巻 2 号 p. 93-101
    発行日: 1997/01/20
    公開日: 2008/12/19
    ジャーナル フリー
    Generally, there are many methods to categorize unknown data in statistics. In many of these methods, we need sample data to determine a border of groups to which these data belong. Moreover, neural networks are also applicable to classify unknown data. Also in leaning process of neural networks, we have to prepare so-called teaching signals, i.e. sample data.
    In this paper, we propose an empirical scheme to organize neural networks for clustering unknown data which belong to certain two groups. In our scheme, a neural network that satisfies an evaluation function without teaching signals are organized. This evaluation function are determined by a histogram of outputs of the neural network. Generally, neural networks map input data distribution to output one. Maximizing the evaluation function means to separate these two output distributions each other. As an organizing mechanism, genetic algorithm is used because of its convergence ability to global maximum. Some numerical results are presented to confirm a feasibility of the scheme.
  • Daehee KANG, Hideki HASHIMOTO, Fumio HARASHIMA
    1997 年 117 巻 2 号 p. 102-109
    発行日: 1997/01/20
    公開日: 2008/12/19
    ジャーナル フリー
    The shortest/optimal path generation is essential for the efficient operation of mobile robot. This paper will present an algorithm for global path planning to a goal with a mobile robot in an known environment. The algorithm makes use of the modified quadtree data structure to model the environment and uses a genetic algorithm to generate a optimal path for the robot to move. Actually the genetic algorithm consists of two stages, the first stage (named a minor league) checks if a chromosome is reachable to goal position or not, and makes the individuals evolve. And, the only reachable chromosome and the best individuals of them are transferred to the second stage (called a major league) and then are evolved. Finally, the best chromosome of individuals in second stage is survived, so that the optimal/shortest path is generated. It is shown that our proposed method can find out a optimal path very quickly through simulation results.
  • Jianming Lu, Muhammad Shafiq, Takashi Yahagi
    1997 年 117 巻 2 号 p. 110-116
    発行日: 1997/01/20
    公開日: 2008/12/19
    ジャーナル フリー
    We present a new method for the self-tuning control (STC) of nonminimum phase discrete-time systems based on the pole-zero placement. We show that the effect of unstable zeros on the output can be canceled approximately. The long division method is used to decompose a polynomial into a stable polynomial and an unstable one. Finally, the results of computer simulation are presented to illustrate the effectiveness of the proposed method.
  • 曽根 良太, 村田 弘志
    1997 年 117 巻 2 号 p. 117-127
    発行日: 1997/01/20
    公開日: 2008/12/19
    ジャーナル フリー
    In this paper, we propose a new design method of the optimal deadbeat servosystem and regulator. The sequence of an optimal deadbeat control signal for the step change of the desired value and optimal deadbeat regulator control signal can be easily obtained by the matrix computation, using sampled data of the step response of controlled process. Then, the deadbeat controller not only for the series compensation but also local feedback compensation can be designed by using sequence of the control signal. The experimental results controlled by microprocessor and numerical examples are presented. The method proposed here is very simple comparing with other methods and it seems to be useful for the practical application.
  • 大浦 邦彦, 花崎 泉, 秋月 影雄
    1997 年 117 巻 2 号 p. 128-135
    発行日: 1997/01/20
    公開日: 2008/12/19
    ジャーナル フリー
    This paper proposes a procedure of estimating orders and delay-time of the process in closed-loop control system. There are identifiability conditions which assure uniqueness and undiasedness of estimates, but they are not available in practice. Because they require the orders and delay-time, which are practically unknown. Therefore, it is primary to estimate the orders and and delay-time of the process in identification of closed-loop system. In this paper, the authors first investigate the influence caused by the closed-loop on the identification, and then, propose the procedure of estimating orders and delay-time taking account of the consiquence that they have in their work.
  • 加地 太一, 大内 東
    1997 年 117 巻 2 号 p. 136-142
    発行日: 1997/01/20
    公開日: 2008/12/19
    ジャーナル フリー
    We think about the problem to assign the elements to an ordered sequence of stations such that the precedence relations are satisfied. At this time, we obtain the best evaluation value decided under a certain constrained condition to the assignment. This kind of problem includes the line balancing problem etc. In addition, this problem can adjust to various problems by changing constrained condition and objective function. These problems can be shown as a problem for sequential partitions of the nodes of a directed acyclic graph into subsets. We especially consider problem for finding a minimum total cost of the cut edge under the restriction of the size of block.
    In this paper, we propose the general framework for sequential partitions of directed acyclic graphs. And we describe an efficient algorithm that can be used to reduce computational requirements and, possibly storage. We estimate that complexity of the algorithm is the polynomial order, if structure of directed acyclic graphs is near parallel.
  • 小川 哲, 渡辺 隆男, 安田 恵一郎, 横山 隆一
    1997 年 117 巻 2 号 p. 143-149
    発行日: 1997/01/20
    公開日: 2008/12/19
    ジャーナル フリー
    This paper proposes a new algorithm for advancing the generalization ability of multilayer neural networks. The proposed algorithm, based on regularization theory, is a method for determining the regularization parameter, on condition that the training data is shown additionally. It is not a method that solves a problem for all training data again when additional training data is shown, but rather a method that adjusts the regularization parameter to fit additional training data. The characteristics of this algorithm are (1) the prediction error for the additional data is used in evaluating to determine the regularization parameter, (2) Evolution strategies (ES) that is multipoint search method is used for the determination problem of regularization parameter. The evaluation of the regularization parameter varies according to the data added. This study simulated an additional learning problem to examine the performance of the proposed method. And the simulation results are presented in this paper.
  • 荒井 邦晴, 相吉 英太郎
    1997 年 117 巻 2 号 p. 150-157
    発行日: 1997/01/20
    公開日: 2008/12/19
    ジャーナル フリー
    Saccade eye movements are among the most rapid yet precise of all movements produced by higher mammals. Recently we have proposed a spatio-temporal neural network model of the superior colliculus which uses lateral excitatory and inhibitory interconnections to help control both the dynamic and static behavior of saccadic eye movements. In this paper a new learning algorithm integrating genetic algorithms with neural networks for the lateral inhibitory and excitatory interconnections in the saccade generation model is presented. Data base for the training were obtained from neurophysiological experiments, and the training converged well even if random connections were chosen as initial conditions. The resulting network model succeeded in making accurate saccadic eye movements of a variety of sizes while producing realistic spatio-temporal patterns of collicular discharge.
  • 八賀 明
    1997 年 117 巻 2 号 p. 158-167
    発行日: 1997/01/20
    公開日: 2008/12/19
    ジャーナル フリー
    A railway interlocking table is a document describing the functional specification of an interlocking device which is responsible for the safe and efficient train operations. Conventionally, the signal engineer makes the most of his expertise to produce an interlocking table. However, since the expertise involved does not always produce the same interlocking table, beginners may sometimes have difficulty understanding it. In the past, some attempts at automatic generation of interlocking table were made but the core of the expertise has always been left untouched as a hard nut to crack.
    This paper proposes a train route setting principle, based on which two algorithms, “lock algorithm” and “compound algorithm, ” have been established to generate a complete interlocking table. The expertise can successfully realized in these algorithms. The man-made and machine-made interlocking tables are compared to find out their differences. It is concluded that some existing interlocking tables can be improved and the algorithms can contribute to several practical purposes of interlocking table.
  • 鳥飼 孝幸, 武末 高明, 永村 幸博, 久保田 宏一, 工藤 富弘, 赤岩 芳彦
    1997 年 117 巻 2 号 p. 168-178
    発行日: 1997/01/20
    公開日: 2008/12/19
    ジャーナル フリー
    The distribution power line carrier system has been used for an automatic distribution control, in order to advance the highly reliable operation and labor saving on electric power supply in suburban districts. In this system, harmonics noise becomes a serious problem deteriorating the data communication performance.
    This paper proposes a method in which the high speed data communication can be realized under harmonics noise by using DPSK (Differential Phase Shift Keying) and a comb filter. First, the performance of the system was computer-simulated using harmonics noise gathered from in-service power lines. Then, the capability of the system transmission rate to get at 120 bits per second (bps) was confirmed in the field experiment.
  • 園田 敏勝, 梶原 和範, 上田 隆三
    1997 年 117 巻 2 号 p. 179-185
    発行日: 1997/01/20
    公開日: 2008/12/19
    ジャーナル フリー
    This paper proposes a constant temperature bath suitable for improving the performance of the zener diode type reference voltage previously presented. The bath consists of a double-layered Cu-box, which reveals the following distinctive features: (i) The temperature resolution belowes to the level of 10-5°C; (ii) the fluctuation width is smaller than 5×10-5°C (p-p) and (iii) the steady state error remains within 8.6× 10-5°C even if the temperature in the outer box varies by 5.9°C. This bath affirms the reference voltage of sevenfigures resolution.
  • 茂吉 雅典, 横山 清子, 渡辺 與作, 高田 和之
    1997 年 117 巻 2 号 p. 186-191
    発行日: 1997/01/20
    公開日: 2008/12/19
    ジャーナル フリー
    The heart rate variability (HRV) is one of information about physiological workloads. Mean heartrates, coefficients of variance of R-R intervals, power spectrum density functions are used as evaluation parameters of HRV. Power spectrum density function is the expression in frequency domain. Physiological correspondence for frequency components of power spectrum density function is evidence, so it had been examined in each area. The example which response function in time domain was used for is not recognized in evaluation of HRV. In this paper, we aimed at analysis with response function in evaluation of HRV during graded head-up tilt. The response function is estimated from the autoregressive model. The method is simple and easy to obtain. It is not accompanied with difficulty to giving input stimulation to the body from the outside. The Rapid Heart Rate Increase, time constant, settling time and power of the response function are suggested as evaluation parameters of HRV by response function. As a result of our analysis, response function in time domain can be used as the evaluation parameters of HRV.
  • 山田 啓一, 中野 倫明, 山本 新
    1997 年 117 巻 2 号 p. 192-197
    発行日: 1997/01/20
    公開日: 2008/12/19
    ジャーナル フリー
    We have developed a robust vision system for 2D positioning of industrial parts against the change of lighting condition. A wide dynamic range vision sensor, we had developed previously, was used to avoid the saturation of images of the objects. Additionally, a gray scale pattern matching technique was employed for the robust image processing. Performance of the developed system with the wide dynamic range vision sensor was investigated experimentally in comparison with that of the system with a conventional video camera. The probability of correct positioning of an object under a changing lighting condition, simulating that in a factory, resulted in 100% for the wide dynamic range vision sensor while 83% for the conventional video camera. The experimental result has confirmed the effectiveness of the developed system, revealing that the dynamic range expansion of the video cameras is very effective for realizing robust robot vision systems.
  • 内田 雅文, 井出 英人, 横山 修一
    1997 年 117 巻 2 号 p. 198-204
    発行日: 1997/01/20
    公開日: 2008/12/19
    ジャーナル フリー
    In this paper, we study ways to avoid obstacles in case of action of an autonomous robot in a dynamic environment. We considered that a problem of obstacle avoidance result in one of the robustness against environment changes in a motion plan. Supposing that it is able to forecast of a working environment of the future and to plan ahead motion for the future, it is possible that a robot takes action in response to environment changes. As a result, we may hope for an improvement in the robustness of a motion plan.
    This study especially take notice of a path of a mobile robot in a motion plan of an autonomous robot. We considered that a transfer from a pattern of environment information to one of plan is a path plan and studied ways to proccess an occupied region pattern of a working environment in parallel. This paper proposes a technique for to forecast of environment changes and to plan ahead motion for the future.
  • 毛 峡, 星野 勉, 牟田 一彌
    1997 年 117 巻 2 号 p. 205-206
    発行日: 1997/01/20
    公開日: 2008/12/19
    ジャーナル フリー
  • Heywood Absaloms, Takehiko Tomikawa
    1997 年 117 巻 2 号 p. 207-208
    発行日: 1997/01/20
    公開日: 2008/12/19
    ジャーナル フリー
feedback
Top